---
title: "Connecting Tugger to Google Gemini"
slug: "connecting-tugger-to-google-gemini"
updated: 2026-03-02T16:53:36Z
published: 2026-03-02T16:53:36Z
canonical: "docs.tuggerapp.com/connecting-tugger-to-google-gemini"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://docs.tuggerapp.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Connecting Tugger to Google Gemini

This guide walks you through connecting Tugger AI Insights to Google Gemini via the Gemini CLI, so you can ask questions about your business data in natural language from your terminal.

Setup takes around 10 minutes.

> **Note:** Gemini MCP is available through the Gemini CLI (a command line tool), not the Gemini web app at gemini.google.com.

## What you'll need

- A **Tugger account** with at least one active connector that has completed its first sync
- **MCP access enabled by your Tugger admin** - by default, users have no MCP access. Your admin needs to grant you access on the User Permissions tab in [MCP Configuration](/docs/tugger-mcp-configuration) before you can connect.
- **Node.js version 18 or higher** installed on your computer
- A **Google account** to authenticate with Gemini

## Step 1: Find your MCP server URL in Tugger

> **Note:** This step requires administrator access. If you're not an admin, ask your Tugger admin for the server URL.

1. Log in to Tugger at [tuggerapp.com](https://tuggerapp.com)
2. Click **MCP Configuration** in the menu
3. On the **Setup Guides** tab, you'll see your **Server URL** at the top of the page

![Screenshot: MCP Configuration - Setup Guides tab showing Server URL with copy button](https://cdn.document360.io/915cfc9e-45d4-4924-af8d-4cea920ef985/Images/Documentation/mcsetupguides-server-url.jpg)

1. Click the copy button next to the URL to copy it to your clipboard

## Step 2: Install the Gemini CLI

Open a terminal and run the following command to install the Gemini CLI globally:

```
npm install -g @google/gemini-cli
```

Once installed, run Gemini once so it generates its configuration directories:

```
gemini
```

You can close it with **Ctrl+C** (or **Cmd+C** on macOS) after it starts.

## Step 3: Configure the Tugger MCP server

1. Navigate to the Gemini configuration directory:

```
cd ~/.gemini
```

If the `.gemini` directory does not exist, make sure you ran `gemini` at least once in Step 2.

1. Open the `settings.json` file in a text editor. For example:

```
nano settings.json
```

1. Add the Tugger MCP server to the `mcpServers` section. Replace `YOUR_TUGGER_SERVER_URL` with the Server URL you copied from Tugger:

```
{
  "mcpServers": {
    "tugger": {
      "httpUrl": "YOUR_TUGGER_SERVER_URL"
    }
  }
}
```

1. Save the file and close the editor

Make sure the file is valid JSON with no trailing commas or syntax errors.

## Step 4: Authenticate and connect

1. Start Gemini:

```
gemini
```

1. Run the authentication command for Tugger:

```
/mcp auth tugger
```

1. A browser window will open showing the Tugger OAuth login screen. Log in with your Tugger credentials and approve access.

![Screenshot: Tugger OAuth login/authorisation screen](https://cdn.document360.io/915cfc9e-45d4-4924-af8d-4cea920ef985/Images/Documentation/tugger-login.jpg)

1. Once authenticated, you will be returned to the Gemini CLI. To verify the connection, run:

```
/mcp
```

You should see Tugger listed with its available tools.

> **Note:** Your Tugger admin controls which connectors you can access through AI Insights. If you can't see data from a particular connector, ask your admin to check your permissions in [MCP Configuration](/docs/tugger-mcp-configuration).

## Step 5: Start asking questions

You're all set. Start asking about your data directly in the Gemini CLI.

Try one of these to get started:

- *"What connectors do I have access to in Tugger?"* - a good first question to confirm what data is available
- *"Show me a summary of my data"* - lets the AI explore what's there
- *"What are my top 10 customers by revenue?"* - if you have an accounting connector like Xero or QuickBooks

For more example prompts and tips on getting the best answers, see [Tips for Getting the Best Results from AI Insights](/docs/tips-for-getting-the-best-results-from-ai-insights).

## Troubleshooting

**"gemini: command not found"**

- Make sure Node.js version 18 or higher is installed (`node --version` to check).
- Make sure the Gemini CLI installed correctly. Try running `npm install -g @google/gemini-cli` again.

**"/mcp auth tugger" fails or no browser opens**

- Check that the Server URL in your `settings.json` matches the URL shown in Tugger's MCP Configuration page.
- Make sure the `settings.json` file is valid JSON with no syntax errors.

**Connected but no data showing**

- Check that at least one connector has completed its first sync in Tugger.
- Ask your Tugger admin to check your access level on the User Permissions tab in MCP Configuration. By default, users have No Access and need to be granted permission.

**Getting unexpected or incomplete results**

- Try being more specific in your questions - see [Tips for Getting the Best Results from AI Insights](/docs/tips-for-getting-the-best-results-from-ai-insights).
- Make sure the connector containing the data you're asking about is included in your MCP access permissions.

**Need to disconnect or reconnect?**

- In the Gemini CLI: remove or edit the Tugger entry in `~/.gemini/settings.json`, then restart Gemini.
- In Tugger: your admin can deactivate the OAuth client from the OAuth Clients tab in MCP Configuration.

---

**Related articles:**

- [What is Tugger AI Insights?](/docs/what-is-tugger-ai-insights)
- [Tips for Getting the Best Results from AI Insights](/docs/tips-for-getting-the-best-results-from-ai-insights)
- [Tugger MCP Configuration](/docs/tugger-mcp-configuration)